Edge Detection Based Autofocus Algorithm to Detect Accurate Camera Working Distance

Abstract

In the last 10 years, industrial image processing has been used in many areas such as automotive, medicine and aviation. Especially for sensitive measurement applications, camera control systems are preferred due to their speed and quality advantages. In this study, a method is proposed to determine the optimum camera position at the image acquisition phase for industrial cameras. In the proposed approach, images are captured while the camera and telecentric lens, which are connected to the servo motor, are moving up and down. The edge informations of the captured images are extracted in the x and y directions and the amplitude value is calculated. Then, meaningful and meaningless edge information at each motor position is found with a median-based approach. Finally, the sharpness of the images are arranged by statistical analysis and the optimum focal point for the camera is found. Sobel, Robert and Prewitt Filters were applied to determine the best method for extracting edge information. Measurements were performed on the workpieces using camera positions where the focus was optimum for each method. According to the results, it has been observed that the camera position, which produces the closest result to the real values, is obtained from the focus determined by the Sobel Filter. The proposed method is leaded to researchers to provide automatic focus for camera control systems.